Assessment of desertification trends in Sistan Plain, Iran using Rs and GIS

2012 
ABSTRACT Desertification is land degradation and loss of biological productivity in arid, semi-arid and dry sub-humid areas due to climatic variations and destructive human activities. Since the vegetative mass is the major biological production of drylands, one can consider the deterioration of plant cover as the main indicator of desertification process. Sistan plain is an arid region in south-eastern Iran that suffers from severe types of desertification and land degradation. To assess the trend of vegetation cover changes that was occured during 1990-2006 period we conducted a two-years study in the region. The basic tools that employed in this study were in-field sampled data, TM and ETM+ images plus with RS softwares. All needed processes were done including geometric and radiometric corrections, mosaicing, clipping, PCA and Tasseled cap. Field sampling was carried out in 26 sites with nine plots in each site. To determine the best vegetation index (VI), we evaluated 83 VIs, among them WDVI, TSAVI 2 , MSAVI 1 and NDVI ab proved to be the best for separating and discriminating the plant cover on the images with correlation coefficients of 0.89, 0.86, 0.84 and 0.70, respectively. Then by applying an appropriate treshold, the images were classified in two categories: vegetation and non-vegetation. Kappa coefficients of resulted classification maps for four selected indices calculated using confusion matrix, suggesting that WDVI map is the the most accurate map. Finally, change detection map was obtained by subtracting images, showing that 11.37% (95288.9 ha) of the study area experienced changes in vegetation cover from year 1990 to 2006. Of all changed area 89.23% belongs to Desertified class while just 10.77% of changed area indicates the Restoration. KEYWORDS : desertification, trend, vegetation indices, WDVI, Sistan, vegetation cover
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    0
    References
    10
    Citations
    NaN
    KQI
    []